James Kyton writes about what might be expected in the next 2 releases of Windows , which he claims we should start seeing the results from within the next 2 years. There are a few things I wanted to comment on...
First, I have to say his timing with the next release is probably about right; however, with Balmer's intentions on time between releases, the following release shouldn't be any more than 3 years after that. In the same breath, I do question a completely new OS in 5 years time. Then again, all they have to do is rewrite the core and update immediately dependant parts; I imagine they've done a somewhat decent job of separating core functionality. At the same time, I know a complete rewrite would be better.
James claims Fiji will come with .NET 3.5. I think I can say what's being called .NET "3.5" today will be out well before the next version of Winodws. While I'm on the subject, the "3.5" in .NET "3.5" is not the expected version number. Instead, this is merely an abstract identifier for a future version of .NET that will include certain features. While it may not be likely, there could potentially be a release before this one. Think of this as Microsoft sharing information without committing to anything. James does say the .NET release could be 4.0, tho, and I think that could be a possibility. Whatever the version is, expect a new version of .NET with Windows.
While integration with Windows Live is definitely a very likely possibility, expect to see a lot of push-back. All I can say is there better be some sort of provider framework built around these. I've mentioned this before in regards to the search features in Vista. I've thought about looking into replacing some of these capabilities with open source alternatives, but the time just isn't there.
Speech recognition in Windows is a gimme. This is obvious with all three of the latest tool releases: Vista, Exchange, and Office. Looking into these will show a vast array of speech tools. I expect speech to be the next great input device in the coming years. Innovations in this space will be very interesting.
The future of managed code is another area I've talked about a lot in the past. To sum up my thoughts, I see a completely managed OS being the ultimate goal of Microsoft. The idea that all non-managed code will run in a sandbox is a touchy subject. There's been a soft-spoken battle between managed and unmanaged code since .NET came out. Most of this has been because the impression was that C++ developers had no choice but to upgrade to .NET. In fact, that's never been true. That's still the case; even in the future of Windows, if you ask me. James seems to be [inadvertantly] backing up my opinions with his comments. I see this being a huge move in the right direction. I expect managed and unmanaged code to completely switch sides in terms of performance, power, and respect, just to name a few.